Eurythermalism and the temperature dependence ... - Semantic Scholar
The FASEB Journal • Research Communication
Eurythermalism and the temperature dependence of enzyme activity Charles K. Lee,* Roy M. Daniel,*,1 Charis Shepherd,† David Saul,† S. Craig Cary,*,‡ Michael J. Danson,§ Robert Eisenthal,储 and Michelle E. Peterson* *Department of Biological Sciences, University of Waikato, Hamilton, New Zealand; †School of Biological Sciences, University of Auckland, Auckland, New Zealand; ‡College of Marine and Earth Studies, University of Delaware, Lewis, Delaware, USA; §Centre for Extremophile Research, Department of Biology and Biochemistry, and 储Department of Biology and Biochemistry, University of Bath, Bath, UK The “Equilibrium Model” has provided new tools for describing and investigating enzyme thermal adaptation. It has been shown that the effect of temperature on enzyme activity is not only governed by ⌬G‡cat and ⌬G‡inact but also by two new intrinsic parameters, ⌬Heq and Teq, which describe the enthalpy and midpoint, respectively, of a reversible equilibrium between active and inactive (but not denatured) forms of enzyme. Twenty-one enzymes from organisms with a wide range of growth temperatures were characterized using the Equilibrium Model. Statistical analysis indicates that Teq is a better predictor of growth temperature than enzyme stability (⌬G‡inact). As expected from the Equilibrium Model, ⌬Heq correlates with catalytic temperature tolerance of enzymes and thus can be declared the first intrinsic and quantitative measure of enzyme eurythermalism. Other findings shed light on the evolution of psychrophilic and thermophilic enzymes. The findings suggest that the description of the Equilibrium Model of the effect of temperature on enzyme activity applies to all enzymes regardless of their temperature origins and that its associated parameters, ⌬Heq and Teq, are intrinsic and necessary parameters for characterizing the thermal properties of enzymes and their temperature adaptation and evolution.—Lee, C. K., Daniel, R. M., Shepherd, C., Saul, D., Cary, S. C., Danson, M. J., Eisenthal, R., Peterson, M. E. Eurythermalism and the temperature dependence of enzyme activity FASEB J. 21, 1934 –1941 (2007) ABSTRACT
Key Words: enzyme temperature optimum 䡠 Equilibrium Model 䡠 growth temperature 䡠 protein stability 䡠 temperature adaptation One of the most important environmental factors for life is temperature. It affects diffusion, membrane fluidity, nucleic acid stability, salt and gas solubility, and, significantly, the behavior of enzymes. Until recently, two parameters have been used to characterize the effect of temperature on enzymes: firstly, the temperature-related reaction kinetics of the chemical reaction catalyzed by the enzyme, expressible as ⌬G‡cat, the free energy of activation of the catalytic reaction, and 1934
secondly thermal stability, which can be described using ⌬G‡inact, the free energy of activation of the thermal denaturation process. Some functional studies of enzyme temperaturelinked properties have defined a so-called enzyme “temperature optimum” that arises from this two-parameter “Classical Model” (1). However, this “temperature optimum” is not an intrinsic enzyme property, being derived from a complex mixture of both activity and thermal stability effects (1) and dependent on assay duration. Consequently, it is of limited value for measuring enzyme temperature adaptation. Evidence of anomalously low enzyme activity at high temperatures (2, 3) also suggests that the Classical Model is insufficient to fully describe the effect of temperature on enzyme activity. In addition, there has not been any quantitative tool for characterizing an enzyme’s eurythermalism. Traditionally, eurythermalism has been used to describe an organism’s ability to withstand a wide range of temperatures. To date, there is limited information on how eurythermalism is achieved in eurythermal organisms, and an important possibility is that special characteristics of enzymes in these organisms play an important role. Enzyme thermal stability clearly has a general correlation with growth temperature; enzymes from thermophiles are typically more thermally stable than those from mesophiles, and those from mesophiles are generally more stable than their psychrophilic counterparts. However, there are exceptions, such as RNase (4, 5), mammalian adenylate kinase (6, 7), and catalase (8 –12), which are much more stable than might be expected from their environmental temperatures. Factors other than a high temperature environment may in some cases be responsible for high thermal stability. Resistance to proteolysis (13) and organic solvents (14), for example, are known to be associated with high thermal stability. In addition, almost all thermal stabil1
Correspondence: Department of Biological Sciences, University of Waikato, Private Bag 3105, Hamilton 3240, New Zealand. E-mail: [email protected] doi: 10.1096/fj.06-7265com 0892-6638/07/0021-1934 © FASEB
ity determinations are done in the absence of substrate, which often affects stability. Two new, intrinsic, temperature-dependent parameters of enzymes, Teq and ⌬Heq, have recently been described (1) and validated (15). These arise from the Equilibrium Model, which proposes that the active form of enzyme (Eact) is in reversible equilibrium with an inactive form (Einact) and that it is the inactive form that undergoes irreversible thermal inactivation to the thermally denatured state (X), as described in (1).
The behavior of the model is described by kcat, the rate constant of the enzyme-catalyzed reaction, kinact, the rate constant for the irreversible inactivation reaction, and Keq, the equilibrium constant governing the ratio Einact/Eact in the reversible interconversion. The dependence of catalytic rate on temperature is obtained from these constants expressed in terms of the thermodynamic parameters, ⌬G‡cat, ⌬G‡inact, ⌬Heq, and Teq as follows:
description of enzyme temperature-linked behavior, it does so under assay conditions, so that the derived constants such as Teq and ⌬Heq are potentially physiologically based. Since enzyme temperature adaptation is the functional basis of metabolic and organism adaptation studies, the Equilibrium Model is not only useful but necessary for conducting such studies. In this work, the intrinsic temperature-dependent properties of 21 enzymes, including several isoactive sets of enzymes derived from a wide range of growth temperature sources, were determined. Parameters generated from model fitting were analyzed to identify any associations with each other or with growth temperature optima.
MATERIALS AND METHODS Enzymes and reagents For detailed descriptions of enzyme sources, see Supplemental Table 4. All chemicals used in this study were purchased from Sigma-Aldrich Inc. and Wako Chemical Company and are of analytical grade or higher (see Supplemental Table 3). Enzyme assays and data collection
where R is the gas constant, kB is Boltzmann’s constant, and h is Planck’s constant. Teq is the temperature representing the midpoint of transition between the active and inactive (but undenatured) forms of the enzyme. In other words, Teq is the temperature at which Keq ⫽ 1 and ⌬G‡eq ⫽ 0 [we have previously (1) used the term Tm to designate this temperature but now prefer the term Teq (15) since it is now clear that Teq is entirely separated from and unrelated to thermal denaturation]. Of more than 40 enzymes examined in our laboratories (15–17), including those presented here, all followed the temperaturedependent behavior predicted by the Equilibrium Model, and none followed that expected from the Classical Model. Furthermore, Teq is independent of ⌬G‡cat and ⌬G‡inact. Preliminary evidence suggests that the active-inactive transition described by the Equilibrium Model arises from a conformational transition (15) at or near the active site, since the active-inactive equilibrium is established in a timescale at least two orders of magnitude faster than that of thermal denaturation (15); stabilizing and destabilizing agents have no effect on Teq (16, 18); and different substrates can affect Teq without altering stability (16, 18). Not only does the Equilibrium Model provide a TEMPERATURE DEPENDENCE OF ENZYME ACTIVITY
In general terms, these were carried out as described previously (17). All enzyme assays performed in this work were continuous photometric assays measured using a Thermospectronic Helios ␥-spectrophotometer equipped with a Thermospectronic single cell peltier-effect cuvette holder. All reactions were started by the rapid addition and mixing of a few microliters of enzyme solution at zero degrees into 400-3000 l of temperature-equilibrated reaction mixture to enable data collection to begin within the first few seconds; the temperature of the reaction mixture remains within ⫾1°C of desired temperature during data collection. Absorbance data were collected at one-second intervals for three minutes using Vision (version 1.25, Thermo Spectronic) on a Windows PC connected to the Helios, within a range of temperatures depending on the source of the enzyme (see Material and Methods in Supplemental Data for assay details). In all cases, precautions were taken so that no decrease in reaction rates was due to substrate depletion (see Material and Methods in Supplemental Data). Where possible, substrate concentrations were maintained at over ten times Km at all temperatures, otherwise appropriate adjustments were made to compensate for the effect of increases in Km. No signs of substrate/ product inhibition were observed under the experimental conditions described. Blank rates were measured at all temperatures and used to correct reaction rates if they were significant (see Table 3 in Supplemental Data for assay conditions for individual enzymes). Based on the variation between the individual triplicate rates from which the parameters are derived for all the enzymes we have assayed thus far, we find that the experimental errors in the determination of ⌬G‡cat, ⌬G‡inact, and Teq, are ⬍0.5% and ⬍6% in the determination of ⌬Heq (17). Data analysis The method for processing experimental data is as described previously in detail (17). Absorbance data from 1935
Vision were first converted in Excel (version 11 for Windows, Microsoft) to progress curves of product concentration (M). The data were subsequently imported into Scientist (version 2.01, MicroMath), where a set of common estimates of thermodynamic parameters (see Supplemental Data) was optimized by “Simplex” searching through the complete data set (product concentration vs. time vs. temperature). The optimized estimates of parameters were then used as initial values to “Least Square” fit the complete data set to the Equilibrium Model, which generated the final values of the parameters; ⌬G‡cat, ⌬G‡inact, ⌬Heq, and Teq. The generated parameters were then used in the “Zero-time Model” (17) to draw a zero-time activity vs. temperature plot. This plot was used to visually estimate Topt and HWHM. Initial rates estimated from assay data were also plotted on the zero-time plot to verify the validity of the fittings. A standalone MATLAB (Version 7.1.0.246 (R14) Service Pack 3, The MathWorks, Inc.) application, enabling the facile derivation of the Equilibrium Model parameters from a Microsoft Office Excel file of experimental progress curves (product concentration vs. time), is available on CD from the corresponding author. This application is suitable for computers running Microsoft Windows XP and is for noncommercial research purposes only. Protein determination Protein determinations were performed using the Bradford assay (19). Statistical analysis All parameters were imported into STATISTICA (20) for further analysis. Correlation analysis was done for nine parameters, i.e., growth temperature, ⌬G‡cat, ⌬G‡inact, ⌬Heq, Teq, Topt, HWHM, the difference between Teq and growth temperature, and the difference between Teq and Topt. The nonparametric Goodman-Kruskal Gamma test was used, and Gamma statistics and the associated two-tailed P values were obtained. Gamma statistic is equivalent to Kendall Tau in terms of the underlying assumptions and interpretation, but takes tied ranks (e.g., growth temperature) into account. Two-tailed statistical power values were calculated with these conditions: n ⫽ 21, Rho0 ⫽ 0, Alpha ⫽ 0.05, and “Exact” algorithm. Best-subset regression analysis was done using growth temperature as the response (dependent) factor and ⌬G‡cat, ⌬G‡inact, ⌬Heq, Teq, and Topt as predictors. A maxi-
mum step of 100 and a probability of 0.05 were used for stepwise selection, and best subsets measures were calculated using R squared values. Nonparametric t test was carried out using Mann-Whitney U test, and a two-tailed P value was calculated. Scatter plots were drawn using MiniTab (R14 for Windows, Minitab).
RESULTS AND DISCUSSION All enzymes, including those from extremophiles, obeyed the Equilibrium Model and were found to display zero-time activity optima with temperature, characteristic of the model (e.g., Fig. 1 and Fig. 2). It should be noted that given the “fast-start” of the enzyme assays (see Materials and Methods), the decline in activity at zero time at temperatures above Topt is evidence that equilibration of the active-inactive forms of the enzyme (Eact and Einact) takes place over timescales significantly shorter than we can observe, i.e., ⬍1–3 s and that the equilibration is about two orders of magnitude or more faster than the rate of irreversible activity loss as shown by the activity vs. time plots for each temperature in the 3D graphs in Fig. 2 (although it should be noted that the very slow irreversible denaturation for the ASB HK47 psychrophilic alkaline phosphatase above Topt is not typical of the enzymes examined here). The parameters of all enzymes are listed in Table 1. They are derived by fitting assay data to the Equilibrium Model and thus relate to active enzymes in the presence of substrate and cofactor. The exception is the growth temperature optimum of the source organism, which is cited from various sources of reference (see Supplemental Data). Growth temperatures of the source organisms range from 2°C to 75°C. The difference between the source growth temperature and Teq ranges from ⫺6.1 to 52.6°C. Topt is the graphical optimum temperature of the enzyme at time zero (see Fig. 1). In most cases, Teq is within a few degrees of Topt, and for almost all enzymes Teq, (and Topt) are higher than the respective source growth temperature. In the case of Teq, the mean difference is 21.1⫾15.2°C. This is not surprising, since
Figure 1. Zero-time activity plot of enzymes with different eurythermal properties [Escherichia. coli MDH (A); Bacillus subtilis IPMDH (B); Moritella profunda DHFR (C)]. ⌬Heq is the enthalpic change for the transition between active and inactive forms of the enzyme; Topt is the temperature at which enzyme activity at zero-time is highest; Half Width at Half Maximum (HWHM) is the width of the zero-time temperature/activity plot between Topt and the upper temperature at which 50% of maximum activity is exhibited. For enzyme name abbreviations, see Table 1 legend. 1936
Vol. 21
June 2007
The FASEB Journal
LEE ET AL.
Figure 2. 3D plots of enzymes from different thermal origins [Antarctic sea bacterium HK47 – psychrotrophic (A); Bacillus cereus – mesophilic (B); Caldicelulosiruptor sacchrolyticus – thermophilic (C)]. The plots are generated from parameters derived directly from the raw data.
at Teq half of the enzyme activity is unavailable, while at 20°C below Teq most of the enzyme will be in the active form. HWHM is the upper temperature half width of the zero-time temperature/activity peak at 50% activity (the upper temperature half width has been used because the low temperature half of this peak is dominated by ⌬G‡cat) and is included to show its relationship to ⌬Heq (see Fig. 1); a large ⌬Heq leads to a sharp decline in activity with increasing temperature above Topt (and thus a small HWHM), since ⌬Heq is the enthalpic change associated with the reversible conversion of the active to the inactive form of enzyme. There are wide variations in both HWHM (from 4.5 to 44°C) and ⌬Heq (from 86 to 826 kJ 䡠 mol⫺1), and thus considerable
variations in the sensitivity of enzyme activity to temperature above Teq (and Topt). ⌬G‡inact describes the stability of the enzyme under reaction conditions; as expected, the average ⌬G‡inact of thermophilic enzymes tends to be higher, and the average ⌬G‡inact of psychrophilic/psychrotrophic enzymes lower, than that of mesophilic enzymes. 3D plots of one thermophilic, one mesophilic, and one psychrophilic enzyme are shown in Fig. 2A–C. Although the plots are very similar in form, the temperature axes show that the three enzymes have very distinct responses to temperature. Note that for the ASB HK47 alkaline phosphatase (Fig. 2A), ⌬G‡inact has relatively little effect on activity over time at tempera-
TABLE 1.
⌬G ‡cat is the free energy of activation of the catalytic reaction, and ⌬G ‡inact is that of the thermal denaturation process. ⌬Heq is the enthalpic difference between active (Eact) and inactive forms (Einact) of the enzyme. Teq is the temperature at which the concentration of Eact equals that of Einact . Topt and HWHM are both graphically determined from zero-time temperature versus activity plots (see Fig. 1). See Supplemental Data for complete names of organisms. GDH, glutamate dehydrogenase; MDH, malate dehydrogenase; DHFR, dihydrofolate reductase; ASB, Antarctic sea bacterium, AKP, alkaline phosphatase; AAA, aryl acylamidase; ACP, acid phosphatase; GLU, glucosidase; FUM, fumarate hydratase; GGTP, ␥-glutamyl transferase; PAL, phenylalanine ammonia-lyase; IPMDH, isopropylmalate dehydrogenase.
TEMPERATURE DEPENDENCE OF ENZYME ACTIVITY
1937
TABLE 2.
Strong correlations are in color, while weaker ones are in normal typeface. Correlation analysis was done using the nonparametric Gamma test (where Gamma is equivalent to correlation coefficient on scale of ⫺1 to ⫹1). P value is the measure of correlation by chance, while Power value is the probability of rejecting a false null hypothesis. A low Power value (i.e., lower than 0.8) indicates that a correlation is subject to type II errors (false negative); in other words, the sample size is too small to successfully detect a weak correlation between the 2 parameters. For scatter plots, vertical axes correspond to the parameter along the horizontal plane of the table, and horizontal axes correspond to the parameter below. Least squares regression curves in scatter plots were calculated using a linear model with fit intercept. The regression curves are not a mathematically correct representation of nonparametric correlation and are only used to indicate general trends in the data. See main text for a detailed description of statistical analyses used.
tures somewhat above Topt, and physiological activity is therefore very dependent on Teq, but for the Bacillus cereus DHFR (Fig. 2B), activity over time above Topt depends largely on ⌬G‡inact. Correlation analysis A correlation analysis of the parameters in Table 1 is shown in Table 2. Two sets of significant correlations are evident. The strongest correlation set (in red) involves ⌬Heq, HWHM (negatively), and the difference between Teq and Topt. ⌬Heq is the enthalpic change associated with the reversible, temperature driven, interconversion of an enzyme between its active and inactive state. It can be considered as a measure of the sensitivity of an enzyme’s catalytic activity to temperature. ⌬Heq will therefore influence the broadness of zero-time activity vs. temperature plots; a small ⌬Heq will lead to a broader zero-time activity vs. temperature peak, thus giving rise to a strong negative correlation between ⌬Heq and HWHM. ⌬Heq can thus be considered to be a quantitative measure of an enzyme’s ability 1938
Vol. 21
June 2007
to function over a temperature range: a small ⌬Heq indicating that the enzyme will function at relatively high activity over a wide range of temperatures, i.e., behave in an eurythermal manner. Conversely, a large ⌬Heq indicates stenothermal behavior. The positive correlation between ⌬Heq and the difference between Teq and Topt can be explained on the basis that a low ⌬Heq will confer a broader time-zero graphical peak, and thus Topt is likely to be higher than Teq. Compared to the correlation set above, other correlations involving ⌬Heq, HWHM, or the difference between Teq and Topt are much weaker. It may be that ⌬Heq, and hence an enzyme’s ability to function over a wide range of temperatures (e.g., in an environment with large temperature fluctuations), is more likely to be determined by the temperature range or temperature variability of the environment(s) than it is by other thermal properties, although this hypothesis has not been tested here. The second strong correlation set (in green) is that between Teq and Topt, and Teq and growth temperature; and moderate or weak correlations between all three of
The FASEB Journal
LEE ET AL.
growth temperature, Topt, and ⌬G‡inact, and between Teq and ⌬G‡inact. Since Topt can theoretically be calculated from Teq (15), the correlation between these two is not surprising, and the weak correlation of these two with protein stability (⌬G‡inact) simply confirms the well-established general tendency of thermophiles to have more stable proteins than mesophiles. Of the other correlations in this set, the correlation between Teq and optimal growth temperature is particularly significant. It indicates that in a broad sense, Teq is a better indication of an enzyme’s temperature origin than either its stability (⌬G‡inact) or the graphically determined Topt. This is supported by the result from best-subset regression analysis, which identifies Teq as the single best parameter for predicting growth temperature (data not shown). The moderate negative correlation (in blue) between growth temperature and the difference between Teq and growth temperature is interesting. From Table 1, it is evident that at high growth temperatures (growth temperature ⱖ55°C), i.e., in organisms that are clearly thermophilic, there tend to be relatively small differences between Teq and growth temperature (7.3⫾6.6°C); however, for psychrophiles, psychrotrophs, and low-temperature mesophiles (growth temperature ⱕ31°C), they are larger (27.8⫾15.1°). A nonparametric t test between the Teq and growth temperature differences of thermophilic and nonthermophilic enzymes reveals that the two groups are significantly different (P value⫽0.008) in this respect. A previous study using kcat as a measure of temperature optima also suggested that enzymes tend to have optimal activity at temperatures higher than their hosts’ living temperature, and the higher the growth temperature of the organism the narrower the gap becomes (21). There are several possible reasons for this. There may be a poor match between the environmental temperature of thermophiles and their laboratory growth temperature optima; thermophiles are often isolated from sources that are appreciably hotter than their determined growth temperatures. Alternatively, it might be argued that the origin of the current mesophilic microbial life is thermophilic, i.e., that microbial evolution has proceeded down-temperature from a (hyper)thermophilic last common ancestor (22–24), and that the evolution of Teq lags behind optimal growth temperature. Furthermore, depending on how crucial their roles may be in the cell, different enzymes might evolve at different speeds to adapt their Teq to the environment, thereby resulting in the variation in Teq among enzymes from one organism (e.g., for Bos taurus, Table 1). Conclusions based on weak correlations must be speculative. Although its P value is slightly over the adopted threshold of 0.05, the correlation between ⌬G‡cat and ⌬G‡inact may suggest that ⌬G‡cat is higher for more stable proteins and thus support various studies (21, 25, 26) that suggested that a change in activation energy (⌬G‡cat) might be important to the temperature-adaptation of psychrophilic enzymes. To investiTEMPERATURE DEPENDENCE OF ENZYME ACTIVITY
gate this idea, correlation analysis was performed with psychrophilic and psychrotrophic enzymes (growth temperatureⱕ20°C, not including plant enzymes) excluded; the correlation between ⌬G‡inact and ⌬G‡cat was found to deteriorate significantly (P value⬎0.1), tending to confirm that a lower ⌬G‡cat might indeed be one of the main adaptations for psychrophilic enzymes, while nonpsychrophilic enzymes take on other means of adaptation. The scatter plots in Table 2 are consistent with the correlations described. Growth temperature In general, any correlations with growth temperature must be regarded as tentative since determining environmental growth temperature from laboratory-determined optimal growth temperatures of bacteria is problematic, partly because the temperature of most microbial environments varies with time, including that of Escherichia coli, for example (27, 28), and partly because the laboratory environment on which these are usually based may not represent environmental conditions. In the absence of better measures, however, we have generally adopted here the published data for organisms involved in this study (see Supplemental Data). Individual enzymes The DHFR from the strict psychrophile Moritella profunda, isolated from deep Atlantic sediments (29, 30), is a surprisingly stable enzyme with a concomitantly high Teq, while its ⌬Heq (104 kJ䡠mol⫺1) is significantly lower than the 25th percentile (115 kJ䡠mol⫺1), suggesting eurythermal behavior. We speculate that its low temperature activity has been achieved by evolving a lower ⌬Heq into an enzyme with an incidentally high Teq, thus conferring eurythermalism. Part of the lower half (below Topt) of the enzyme activity curve is influenced by ⌬Heq, but the effect of ⌬G‡cat may dominate, as stated above. The alkaline phosphatase from isolate ASB HK47 (Fig. 2A) gives another illustration of the relative effects of the four parameters on enzyme activity over a range of temperatures. For many enzymes (e.g., B. cereus DHFR, Fig. 2B), the denaturation rate is such that it will play the major role in controlling enzyme activity over timescales important to the organism (over minutes rather than seconds) at temperatures above Teq. However, for the ASB HK47 alkaline phosphatase, it is evident that the lower rate of denaturation above Teq leads to a situation where activity will be dominated by Teq and ⌬Heq over significant timescales. To achieve temperature adaptation, evolution imposes a strong selection force for thermophilic enzymes to have both high thermal stability and a high Teq. High thermal stability has been thought to be achieved at the expense of flexibility to explain why thermophilic enzymes generally exhibit relatively little activity at lower temperatures (31). However, the rigidity required for 1939
stability is global, while the flexibility required for catalysis is likely to be local (32), as demonstrated by some exceptions (33, 34). It has been suggested that a combination of local flexibility at the active site and high overall stability (35, 36), is behind such phenomena. In other words, enzymes active at low temperatures do not necessarily have to be thermally unstable (37, 38). Recent reports also suggested that the higher flexibility of psychrophilic enzymes is local and on the microsecond to millisecond timescale (39), while their global flexibility may not be necessarily higher (38). In addition, random and directed mutagenesis studies (37, 40, 41) have shown that it is possible to engineer an enzyme so that it retains low-temperature activity while gaining thermostability.
grant from the New Zealand Marsden Fund. We thank Y. Xu at Free University of Brussels for M. profunda DHFR, C. Monk for providing purified Caldicellulosiruptor saccharolyticus -GLU, D. Clement for providing purified Geobacillus stearothermophilus DHFR, and J. Klinman at UC Berkeley for the gift of the G. stearothermophilus DHFR clone. We also thank Dr. Ray Littler at the University of Waikato for some assistance with the statistical analysis. We are grateful to O. Planckaert and A.-C. Tsuei for help with data collection.
REFERENCES 1. 2.
CONCLUSION
3.
The correlations observed here do not necessarily denote a causative relationship. The finding that Teq correlates best with growth temperature does not necessarily mean that it is the parameter most directly selectable by evolution, although it does provide insight into the effect of temperature on the evolution of enzymes. It should also be noted that the strong correlation between Teq and growth temperature means that the correlation is likely to be reproducible with any selection of enzymes of similar number and diversity to this study. ⌬Heq is a new quantitative measure to characterize eurythermalism at an enzymatic level. The new parameters (i.e., Teq and ⌬Heq), combined with established parameters (i.e., ⌬G‡cat and ⌬G‡inact), define fully how temperature influences enzyme activity, while the former provide new tools to describe and investigate this influence. Much of the data presented here comes from bacterial sources, although results from correlation analysis remain similar when data from eukaryotic sources are excluded (data not shown). Although the 21 sets of data from discrete enzymes presented here are a relatively small data set, the main correlations are robust and generally consistent with findings or proposals based on other data; they also provide new targets and perspectives for molecular and structural studies. The development and verification of the Equilibrium Model have provided additional and quantitative measures of the thermal behavior of enzymes (1, 15) that are essential for describing the effect of temperature on enzyme activity and useful parameters for measuring the temperature adaptation of enzymes. Additionally, the results here show the Equilibrium Model’s potential usefulness as a tool to investigate the relationship between enzyme thermal properties and the influence of temperature on the physiology and evolution of the host organism. This work was partly supported by a grant from the National Science Foundation (Biocomplexity 0120648) and a 1940
Vol. 21
June 2007
4. 5.
6.
7. 8.
9. 10. 11.
12. 13. 14.
15.
16. 17.
Daniel, R. M., Danson, M. J., and Eisenthal, R. (2001) The temperature optima of enzymes: a new perspective on an old phenomenon. Trends Biochem. Sci. 26, 223–225 Thomas, T. M., and Scopes, R. K. (1998) The effects of temperature on the kinetics and stability of mesophilic and thermophilic 3-phosphoglycerate kinases. Biochem. J. 330, 1087–1095 Gerike, U., Danson, M. J., Russell, N. J., and Hough, D. W. (1997) Sequencing and expression of the gene encoding a cold-active citrate synthase from an Antarctic bacterium, strain DS2–3R. Eur. J. Biochem. 248, 49 –57 Notomista, E., Catanzano, F., Graziano, G., Dal Piaz, F., Barone, G., D’Alessio, G., and Di Donato, A. (2000) Onconase: an unusually stable protein. Biochemistry 39, 8711– 8718 Padmanabhan, S., Zhou, K., Chu, C. Y., Lim, R. W., and Lim, L. W. (2001) Overexpression, biophysical characterization, and crystallization of ribonuclease I from Escherichia coli, a broadspecificity enzyme in the RNase T2 family. Arch. Biochem. Biophys. 390, 42–50 Noda, L., and Kuby, S. A. (1957) Adenosine triphosphateadenosine monophosphate transphosphorylase (myokinase). I. Isolation of the crystalline enzyme from rabbit skeletal muscle. J. Biol. Chem. 226, 541–549 Tsuboi, K. K., and Chervenka, C. H. (1975) Adenylate kinase of human erythrocyte. Isolation and properties of the predominant inherited form. J. Biol. Chem. 250, 132–140 Diaz, A., Rangel, P., Montes de Oca, Y., Lledias, F., and Hansberg, W. (2001) Molecular and kinetic study of catalase-1, a durable large catalase of Neurospora crassa. Free. Radic. Biol. Med. 31, 1323–1333 Calera, J. A., Sanchez-Weatherby, J., Lopez-Medrano, R., and Leal, F. (2000) Distinctive properties of the catalase B of Aspergillus nidulans. FEBS Lett. 475, 117–120 Wasserman, B. P., and Hultin, H. O. (1981) Effect of deglycosylation on the stability of Aspergillus niger catalase. Arch. Biochem. Biophys. 212, 385–392 Lopez-Medrano, R., Ovejero, M. C., Calera, J. A., Puente, P., and Leal, F. (1995) An immunodominant 90-kilodalton Aspergillus fumigatus antigen is the subunit of a catalase. Infect. Immun. 63, 4774 – 4780 Switala, J., O’Neil, J. O., and Loewen, P. C. (1999) Catalase HPII from Escherichia coli exhibits enhanced resistance to denaturation. Biochemistry 38, 3895–3901 Daniel, R. M., Cowan, D. A., Morgan, H. W., and Curran, M. P. (1982) A correlation between protein thermostability and resistance to proteolysis. Biochem. J. 207, 641– 644 Owusu, R. K., and Cowan, D. A. (1989) Correlation between microbial protein thermostability and resistance to denaturation in aqueous: organic solvent two-phase systems. Enzyme Microbial Techn. 11, 568 Peterson, M. E., Eisenthal, R., Danson, M. J., Spence, A., and Daniel, R. M. (2004) A new intrinsic thermal parameter for enzymes reveals true temperature optima. J. Biol. Chem. 279, 20717–20722 Eisenthal, R., Peterson, M. E., Daniel, R. M., Danson, M. J. (2006) The thermal behaviour of enzymes: implications for biotechnology. Trends Bio/Tech. 24, 289 –292 Peterson, M. E., Daniel, R. M., Danson, M. J., and Eisenthal, R. (2007) The dependence of enzyme activity on temperature: determination and validation of parameters. Biochem. J. 402, 331–337.
The FASEB Journal
LEE ET AL.
18. 19. 20. 21.
22. 23. 24. 25.
26.
27. 28. 29.
30.
31.
Peterson, M. E. (2005) Evidence for a third thermal parameter of enzymes. PhD Thesis: University of Waikato, Hamilton, New Zealand Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248 –254 StatSoft Inc (2005) STATISTICA, version 7.1. p. www.statsoft. com, Tulsa, OK 74104 U. S. A. Georlette, D., Blaise, V., Collins, T., D’Amico, S., Gratia, E., Hoyoux, A., Marx, J. C., Sonan, G., Feller, G., and Gerday, C. (2004) Some like it cold: biocatalysis at low temperatures. FEMS Microbiol. Rev. 28, 25– 42 Schwartzman, D. W., and Lineweaver, C. H. (2004) The hyperthermophilic origin of life revisited. Biochem. Soc. Trans. 32, 168 –171 Pace, N. R. (1997) A molecular view of microbial diversity and the biosphere. Science 276, 734 –740 Wachtershauser, G. (1998) Thermophiles: the Keys to Molecular Evolution and the Origin of Life? (Wiegel, K., Adams, M., ed) pp. 47–57, Taylor and Francis, Philadelphia D’Amico, S., Claverie, P., Collins, T., Georlette, D., Gratia, E., Hoyoux, A., Meuwis, M. A., Feller, G., and Gerday, C. (2002) Molecular basis of cold adaptation. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 357, 917–925 Johns, G. C., and Somero, G. N. (2004) Evolutionary convergence in adaptation of proteins to temperature: A4-lactate dehydrogenases of Pacific damselfishes (Chromis spp.). Mol. Biol. Evol. 21, 314 –320 Flint, K. P. (1987) The long-term survival of Escherichia coli in river water. J. Appl. Bacteriol. 63, 261–270 Hartl, D. L., and Dykhuizen, D. E. (1984) The. population. genetics. of. Escherichia. coli. Annu. Rev. Genet. 18, 31– 68 Xu, Y., Nogi, Y., Kato, C., Liang, Z., Ruger, H. J., De Kegel, D., and Glansdorff, N. (2003) Moritella profunda sp. nov., and Moritella abyssi sp. nov., two psychropiezophilic organisms isolated from deep Atlantic sediments. Int. J. Syst. Evol. Microbiol. 53, 533–538 Xu, Y., Feller, G., Gerday, C., and Glansdorff, N. Moritella cold-active dihydrofolate reductase: are there natural limits to optimization of catalytic efficiency at low temperature? (2003) J. Bacteriol. 185, 5519 –5526 Ichikawa, J. K., and Clarke, S. (1998) A highly active protein repair enzyme from an extreme thermophile: the L-isoaspartyl methyltransferase from Thermotoga maritima. Arch. Biochem. Biophys. 358, 222–231
TEMPERATURE DEPENDENCE OF ENZYME ACTIVITY
32. 33.
34.
35.
36. 37. 38.
39.
40.
41.
Daniel, R. M., Dunn, R. V., Finney, J. L., and Smith, J. C. (2003) The role of dynamics in enzyme activity. Annu. Rev. Biophys. Biomol. Struct. 32, 69 –92 Merz, A., Knochel, T., Jansonius, J. N., and Kirschner, K. (1999) The hyperthermostable indoleglycerol phosphate synthase from Thermotoga maritima is destabilized by mutational disruption of two solvent-exposed salt bridges. J. Mol. Biol. 288, 753–763 Pappenberger, G., Schurig, H., and Jaenicke, R. (1997) Disruption of an ionic network leads to accelerated thermal denaturation of D-glyceraldehyde-3-phosphate dehydrogenase from the hyperthermophilic bacterium Thermotoga maritima. J. Mol. Biol. 274, 676 Aguilar, C. F., Sanderson, I., Moracci, M., Ciaramella, M., Nucci, R., Rossi, M., and Pearl, L. H. (1997) Crystal structure of the beta-glycosidase from the hyperthermophilic archeon Sulfolobus solfataricus: resilience as a key factor in thermostability. J. Mol. Biol. 271, 789 – 802 Daniel, R. M., Dines, M., and Petach, H. H. (1996) The denaturation and degradation of stable enzymes at high temperatures. Biochem. J. 317, 1–11 Van den Burg, B., Vriend, G., Veltman, O. R., Venema, G., and Eijsink, V. G. (1998) Engineering an enzyme to resist boiling. Proc. Natl. Acad. Sci. U. S. A. 95, 2056 –2060 Svingor, A., Kardos, J., Hajdu, I., Nemeth, A., and Zavodszky, P. (2001) A better enzyme to cope with cold. Comparative flexibility studies on psychrotrophic, mesophilic, and thermophilic IPMDHs. J. Biol. Chem. 276, 28121–28125 Georlette, D., Damien, B., Blaise, V., Depiereux, E., Uversky, V. N., Gerday, C., and Feller, G. (2003) Structural and functional adaptations to extreme temperatures in psychrophilic, mesophilic, and thermophilic DNA ligases. J. Biol. Chem. 278, 37015–37023 Gershenson, A., Schauerte, J. A., Giver, L., and Arnold, F. H. (2000) Tryptophan phosphorescence study of enzyme flexibility and unfolding in laboratory-evolved thermostable esterases. Biochemistry 39, 4658 – 4665 Miyazaki, K., Wintrode, P. L., Grayling, R. A., Rubingh, D. N., and Arnold, F. H. (2000) Directed evolution study of temperature adaptation in a psychrophilic enzyme. J. Mol. Biol. 297, 1015–1026 Received for publication September 12, 2006. Accepted for publication January 18, 2007.
1941
Eurythermalism and the temperature dependence of enzyme activity Charles K. Lee,* Roy M. Daniel,*,1 Charis Shepherd,† David Saul,† S. Craig Cary,*,‡ Michael J. Danson,§ Robert Eisenthal,储 and Michelle E. Peterson* *Department of Biological Sciences, University of Waikato, Hamilton, New Zealand; †School of Biological Sciences, University of Auckland, Auckland, New Zealand; ‡College of Marine and Earth Studies, University of Delaware, Lewis, Delaware, USA; §Centre for Extremophile Research, Department of Biology and Biochemistry, and 储Department of Biology and Biochemistry, University of Bath, Bath, UK The “Equilibrium Model” has provided new tools for describing and investigating enzyme thermal adaptation. It has been shown that the effect of temperature on enzyme activity is not only governed by ⌬G‡cat and ⌬G‡inact but also by two new intrinsic parameters, ⌬Heq and Teq, which describe the enthalpy and midpoint, respectively, of a reversible equilibrium between active and inactive (but not denatured) forms of enzyme. Twenty-one enzymes from organisms with a wide range of growth temperatures were characterized using the Equilibrium Model. Statistical analysis indicates that Teq is a better predictor of growth temperature than enzyme stability (⌬G‡inact). As expected from the Equilibrium Model, ⌬Heq correlates with catalytic temperature tolerance of enzymes and thus can be declared the first intrinsic and quantitative measure of enzyme eurythermalism. Other findings shed light on the evolution of psychrophilic and thermophilic enzymes. The findings suggest that the description of the Equilibrium Model of the effect of temperature on enzyme activity applies to all enzymes regardless of their temperature origins and that its associated parameters, ⌬Heq and Teq, are intrinsic and necessary parameters for characterizing the thermal properties of enzymes and their temperature adaptation and evolution.—Lee, C. K., Daniel, R. M., Shepherd, C., Saul, D., Cary, S. C., Danson, M. J., Eisenthal, R., Peterson, M. E. Eurythermalism and the temperature dependence of enzyme activity FASEB J. 21, 1934 –1941 (2007) ABSTRACT
Key Words: enzyme temperature optimum 䡠 Equilibrium Model 䡠 growth temperature 䡠 protein stability 䡠 temperature adaptation One of the most important environmental factors for life is temperature. It affects diffusion, membrane fluidity, nucleic acid stability, salt and gas solubility, and, significantly, the behavior of enzymes. Until recently, two parameters have been used to characterize the effect of temperature on enzymes: firstly, the temperature-related reaction kinetics of the chemical reaction catalyzed by the enzyme, expressible as ⌬G‡cat, the free energy of activation of the catalytic reaction, and 1934
secondly thermal stability, which can be described using ⌬G‡inact, the free energy of activation of the thermal denaturation process. Some functional studies of enzyme temperaturelinked properties have defined a so-called enzyme “temperature optimum” that arises from this two-parameter “Classical Model” (1). However, this “temperature optimum” is not an intrinsic enzyme property, being derived from a complex mixture of both activity and thermal stability effects (1) and dependent on assay duration. Consequently, it is of limited value for measuring enzyme temperature adaptation. Evidence of anomalously low enzyme activity at high temperatures (2, 3) also suggests that the Classical Model is insufficient to fully describe the effect of temperature on enzyme activity. In addition, there has not been any quantitative tool for characterizing an enzyme’s eurythermalism. Traditionally, eurythermalism has been used to describe an organism’s ability to withstand a wide range of temperatures. To date, there is limited information on how eurythermalism is achieved in eurythermal organisms, and an important possibility is that special characteristics of enzymes in these organisms play an important role. Enzyme thermal stability clearly has a general correlation with growth temperature; enzymes from thermophiles are typically more thermally stable than those from mesophiles, and those from mesophiles are generally more stable than their psychrophilic counterparts. However, there are exceptions, such as RNase (4, 5), mammalian adenylate kinase (6, 7), and catalase (8 –12), which are much more stable than might be expected from their environmental temperatures. Factors other than a high temperature environment may in some cases be responsible for high thermal stability. Resistance to proteolysis (13) and organic solvents (14), for example, are known to be associated with high thermal stability. In addition, almost all thermal stabil1
Correspondence: Department of Biological Sciences, University of Waikato, Private Bag 3105, Hamilton 3240, New Zealand. E-mail: [email protected] doi: 10.1096/fj.06-7265com 0892-6638/07/0021-1934 © FASEB
ity determinations are done in the absence of substrate, which often affects stability. Two new, intrinsic, temperature-dependent parameters of enzymes, Teq and ⌬Heq, have recently been described (1) and validated (15). These arise from the Equilibrium Model, which proposes that the active form of enzyme (Eact) is in reversible equilibrium with an inactive form (Einact) and that it is the inactive form that undergoes irreversible thermal inactivation to the thermally denatured state (X), as described in (1).
The behavior of the model is described by kcat, the rate constant of the enzyme-catalyzed reaction, kinact, the rate constant for the irreversible inactivation reaction, and Keq, the equilibrium constant governing the ratio Einact/Eact in the reversible interconversion. The dependence of catalytic rate on temperature is obtained from these constants expressed in terms of the thermodynamic parameters, ⌬G‡cat, ⌬G‡inact, ⌬Heq, and Teq as follows:
description of enzyme temperature-linked behavior, it does so under assay conditions, so that the derived constants such as Teq and ⌬Heq are potentially physiologically based. Since enzyme temperature adaptation is the functional basis of metabolic and organism adaptation studies, the Equilibrium Model is not only useful but necessary for conducting such studies. In this work, the intrinsic temperature-dependent properties of 21 enzymes, including several isoactive sets of enzymes derived from a wide range of growth temperature sources, were determined. Parameters generated from model fitting were analyzed to identify any associations with each other or with growth temperature optima.
MATERIALS AND METHODS Enzymes and reagents For detailed descriptions of enzyme sources, see Supplemental Table 4. All chemicals used in this study were purchased from Sigma-Aldrich Inc. and Wako Chemical Company and are of analytical grade or higher (see Supplemental Table 3). Enzyme assays and data collection
where R is the gas constant, kB is Boltzmann’s constant, and h is Planck’s constant. Teq is the temperature representing the midpoint of transition between the active and inactive (but undenatured) forms of the enzyme. In other words, Teq is the temperature at which Keq ⫽ 1 and ⌬G‡eq ⫽ 0 [we have previously (1) used the term Tm to designate this temperature but now prefer the term Teq (15) since it is now clear that Teq is entirely separated from and unrelated to thermal denaturation]. Of more than 40 enzymes examined in our laboratories (15–17), including those presented here, all followed the temperaturedependent behavior predicted by the Equilibrium Model, and none followed that expected from the Classical Model. Furthermore, Teq is independent of ⌬G‡cat and ⌬G‡inact. Preliminary evidence suggests that the active-inactive transition described by the Equilibrium Model arises from a conformational transition (15) at or near the active site, since the active-inactive equilibrium is established in a timescale at least two orders of magnitude faster than that of thermal denaturation (15); stabilizing and destabilizing agents have no effect on Teq (16, 18); and different substrates can affect Teq without altering stability (16, 18). Not only does the Equilibrium Model provide a TEMPERATURE DEPENDENCE OF ENZYME ACTIVITY
In general terms, these were carried out as described previously (17). All enzyme assays performed in this work were continuous photometric assays measured using a Thermospectronic Helios ␥-spectrophotometer equipped with a Thermospectronic single cell peltier-effect cuvette holder. All reactions were started by the rapid addition and mixing of a few microliters of enzyme solution at zero degrees into 400-3000 l of temperature-equilibrated reaction mixture to enable data collection to begin within the first few seconds; the temperature of the reaction mixture remains within ⫾1°C of desired temperature during data collection. Absorbance data were collected at one-second intervals for three minutes using Vision (version 1.25, Thermo Spectronic) on a Windows PC connected to the Helios, within a range of temperatures depending on the source of the enzyme (see Material and Methods in Supplemental Data for assay details). In all cases, precautions were taken so that no decrease in reaction rates was due to substrate depletion (see Material and Methods in Supplemental Data). Where possible, substrate concentrations were maintained at over ten times Km at all temperatures, otherwise appropriate adjustments were made to compensate for the effect of increases in Km. No signs of substrate/ product inhibition were observed under the experimental conditions described. Blank rates were measured at all temperatures and used to correct reaction rates if they were significant (see Table 3 in Supplemental Data for assay conditions for individual enzymes). Based on the variation between the individual triplicate rates from which the parameters are derived for all the enzymes we have assayed thus far, we find that the experimental errors in the determination of ⌬G‡cat, ⌬G‡inact, and Teq, are ⬍0.5% and ⬍6% in the determination of ⌬Heq (17). Data analysis The method for processing experimental data is as described previously in detail (17). Absorbance data from 1935
Vision were first converted in Excel (version 11 for Windows, Microsoft) to progress curves of product concentration (M). The data were subsequently imported into Scientist (version 2.01, MicroMath), where a set of common estimates of thermodynamic parameters (see Supplemental Data) was optimized by “Simplex” searching through the complete data set (product concentration vs. time vs. temperature). The optimized estimates of parameters were then used as initial values to “Least Square” fit the complete data set to the Equilibrium Model, which generated the final values of the parameters; ⌬G‡cat, ⌬G‡inact, ⌬Heq, and Teq. The generated parameters were then used in the “Zero-time Model” (17) to draw a zero-time activity vs. temperature plot. This plot was used to visually estimate Topt and HWHM. Initial rates estimated from assay data were also plotted on the zero-time plot to verify the validity of the fittings. A standalone MATLAB (Version 7.1.0.246 (R14) Service Pack 3, The MathWorks, Inc.) application, enabling the facile derivation of the Equilibrium Model parameters from a Microsoft Office Excel file of experimental progress curves (product concentration vs. time), is available on CD from the corresponding author. This application is suitable for computers running Microsoft Windows XP and is for noncommercial research purposes only. Protein determination Protein determinations were performed using the Bradford assay (19). Statistical analysis All parameters were imported into STATISTICA (20) for further analysis. Correlation analysis was done for nine parameters, i.e., growth temperature, ⌬G‡cat, ⌬G‡inact, ⌬Heq, Teq, Topt, HWHM, the difference between Teq and growth temperature, and the difference between Teq and Topt. The nonparametric Goodman-Kruskal Gamma test was used, and Gamma statistics and the associated two-tailed P values were obtained. Gamma statistic is equivalent to Kendall Tau in terms of the underlying assumptions and interpretation, but takes tied ranks (e.g., growth temperature) into account. Two-tailed statistical power values were calculated with these conditions: n ⫽ 21, Rho0 ⫽ 0, Alpha ⫽ 0.05, and “Exact” algorithm. Best-subset regression analysis was done using growth temperature as the response (dependent) factor and ⌬G‡cat, ⌬G‡inact, ⌬Heq, Teq, and Topt as predictors. A maxi-
mum step of 100 and a probability of 0.05 were used for stepwise selection, and best subsets measures were calculated using R squared values. Nonparametric t test was carried out using Mann-Whitney U test, and a two-tailed P value was calculated. Scatter plots were drawn using MiniTab (R14 for Windows, Minitab).
RESULTS AND DISCUSSION All enzymes, including those from extremophiles, obeyed the Equilibrium Model and were found to display zero-time activity optima with temperature, characteristic of the model (e.g., Fig. 1 and Fig. 2). It should be noted that given the “fast-start” of the enzyme assays (see Materials and Methods), the decline in activity at zero time at temperatures above Topt is evidence that equilibration of the active-inactive forms of the enzyme (Eact and Einact) takes place over timescales significantly shorter than we can observe, i.e., ⬍1–3 s and that the equilibration is about two orders of magnitude or more faster than the rate of irreversible activity loss as shown by the activity vs. time plots for each temperature in the 3D graphs in Fig. 2 (although it should be noted that the very slow irreversible denaturation for the ASB HK47 psychrophilic alkaline phosphatase above Topt is not typical of the enzymes examined here). The parameters of all enzymes are listed in Table 1. They are derived by fitting assay data to the Equilibrium Model and thus relate to active enzymes in the presence of substrate and cofactor. The exception is the growth temperature optimum of the source organism, which is cited from various sources of reference (see Supplemental Data). Growth temperatures of the source organisms range from 2°C to 75°C. The difference between the source growth temperature and Teq ranges from ⫺6.1 to 52.6°C. Topt is the graphical optimum temperature of the enzyme at time zero (see Fig. 1). In most cases, Teq is within a few degrees of Topt, and for almost all enzymes Teq, (and Topt) are higher than the respective source growth temperature. In the case of Teq, the mean difference is 21.1⫾15.2°C. This is not surprising, since
Figure 1. Zero-time activity plot of enzymes with different eurythermal properties [Escherichia. coli MDH (A); Bacillus subtilis IPMDH (B); Moritella profunda DHFR (C)]. ⌬Heq is the enthalpic change for the transition between active and inactive forms of the enzyme; Topt is the temperature at which enzyme activity at zero-time is highest; Half Width at Half Maximum (HWHM) is the width of the zero-time temperature/activity plot between Topt and the upper temperature at which 50% of maximum activity is exhibited. For enzyme name abbreviations, see Table 1 legend. 1936
Vol. 21
June 2007
The FASEB Journal
LEE ET AL.
Figure 2. 3D plots of enzymes from different thermal origins [Antarctic sea bacterium HK47 – psychrotrophic (A); Bacillus cereus – mesophilic (B); Caldicelulosiruptor sacchrolyticus – thermophilic (C)]. The plots are generated from parameters derived directly from the raw data.
at Teq half of the enzyme activity is unavailable, while at 20°C below Teq most of the enzyme will be in the active form. HWHM is the upper temperature half width of the zero-time temperature/activity peak at 50% activity (the upper temperature half width has been used because the low temperature half of this peak is dominated by ⌬G‡cat) and is included to show its relationship to ⌬Heq (see Fig. 1); a large ⌬Heq leads to a sharp decline in activity with increasing temperature above Topt (and thus a small HWHM), since ⌬Heq is the enthalpic change associated with the reversible conversion of the active to the inactive form of enzyme. There are wide variations in both HWHM (from 4.5 to 44°C) and ⌬Heq (from 86 to 826 kJ 䡠 mol⫺1), and thus considerable
variations in the sensitivity of enzyme activity to temperature above Teq (and Topt). ⌬G‡inact describes the stability of the enzyme under reaction conditions; as expected, the average ⌬G‡inact of thermophilic enzymes tends to be higher, and the average ⌬G‡inact of psychrophilic/psychrotrophic enzymes lower, than that of mesophilic enzymes. 3D plots of one thermophilic, one mesophilic, and one psychrophilic enzyme are shown in Fig. 2A–C. Although the plots are very similar in form, the temperature axes show that the three enzymes have very distinct responses to temperature. Note that for the ASB HK47 alkaline phosphatase (Fig. 2A), ⌬G‡inact has relatively little effect on activity over time at tempera-
TABLE 1.
⌬G ‡cat is the free energy of activation of the catalytic reaction, and ⌬G ‡inact is that of the thermal denaturation process. ⌬Heq is the enthalpic difference between active (Eact) and inactive forms (Einact) of the enzyme. Teq is the temperature at which the concentration of Eact equals that of Einact . Topt and HWHM are both graphically determined from zero-time temperature versus activity plots (see Fig. 1). See Supplemental Data for complete names of organisms. GDH, glutamate dehydrogenase; MDH, malate dehydrogenase; DHFR, dihydrofolate reductase; ASB, Antarctic sea bacterium, AKP, alkaline phosphatase; AAA, aryl acylamidase; ACP, acid phosphatase; GLU, glucosidase; FUM, fumarate hydratase; GGTP, ␥-glutamyl transferase; PAL, phenylalanine ammonia-lyase; IPMDH, isopropylmalate dehydrogenase.
TEMPERATURE DEPENDENCE OF ENZYME ACTIVITY
1937
TABLE 2.
Strong correlations are in color, while weaker ones are in normal typeface. Correlation analysis was done using the nonparametric Gamma test (where Gamma is equivalent to correlation coefficient on scale of ⫺1 to ⫹1). P value is the measure of correlation by chance, while Power value is the probability of rejecting a false null hypothesis. A low Power value (i.e., lower than 0.8) indicates that a correlation is subject to type II errors (false negative); in other words, the sample size is too small to successfully detect a weak correlation between the 2 parameters. For scatter plots, vertical axes correspond to the parameter along the horizontal plane of the table, and horizontal axes correspond to the parameter below. Least squares regression curves in scatter plots were calculated using a linear model with fit intercept. The regression curves are not a mathematically correct representation of nonparametric correlation and are only used to indicate general trends in the data. See main text for a detailed description of statistical analyses used.
tures somewhat above Topt, and physiological activity is therefore very dependent on Teq, but for the Bacillus cereus DHFR (Fig. 2B), activity over time above Topt depends largely on ⌬G‡inact. Correlation analysis A correlation analysis of the parameters in Table 1 is shown in Table 2. Two sets of significant correlations are evident. The strongest correlation set (in red) involves ⌬Heq, HWHM (negatively), and the difference between Teq and Topt. ⌬Heq is the enthalpic change associated with the reversible, temperature driven, interconversion of an enzyme between its active and inactive state. It can be considered as a measure of the sensitivity of an enzyme’s catalytic activity to temperature. ⌬Heq will therefore influence the broadness of zero-time activity vs. temperature plots; a small ⌬Heq will lead to a broader zero-time activity vs. temperature peak, thus giving rise to a strong negative correlation between ⌬Heq and HWHM. ⌬Heq can thus be considered to be a quantitative measure of an enzyme’s ability 1938
Vol. 21
June 2007
to function over a temperature range: a small ⌬Heq indicating that the enzyme will function at relatively high activity over a wide range of temperatures, i.e., behave in an eurythermal manner. Conversely, a large ⌬Heq indicates stenothermal behavior. The positive correlation between ⌬Heq and the difference between Teq and Topt can be explained on the basis that a low ⌬Heq will confer a broader time-zero graphical peak, and thus Topt is likely to be higher than Teq. Compared to the correlation set above, other correlations involving ⌬Heq, HWHM, or the difference between Teq and Topt are much weaker. It may be that ⌬Heq, and hence an enzyme’s ability to function over a wide range of temperatures (e.g., in an environment with large temperature fluctuations), is more likely to be determined by the temperature range or temperature variability of the environment(s) than it is by other thermal properties, although this hypothesis has not been tested here. The second strong correlation set (in green) is that between Teq and Topt, and Teq and growth temperature; and moderate or weak correlations between all three of
The FASEB Journal
LEE ET AL.
growth temperature, Topt, and ⌬G‡inact, and between Teq and ⌬G‡inact. Since Topt can theoretically be calculated from Teq (15), the correlation between these two is not surprising, and the weak correlation of these two with protein stability (⌬G‡inact) simply confirms the well-established general tendency of thermophiles to have more stable proteins than mesophiles. Of the other correlations in this set, the correlation between Teq and optimal growth temperature is particularly significant. It indicates that in a broad sense, Teq is a better indication of an enzyme’s temperature origin than either its stability (⌬G‡inact) or the graphically determined Topt. This is supported by the result from best-subset regression analysis, which identifies Teq as the single best parameter for predicting growth temperature (data not shown). The moderate negative correlation (in blue) between growth temperature and the difference between Teq and growth temperature is interesting. From Table 1, it is evident that at high growth temperatures (growth temperature ⱖ55°C), i.e., in organisms that are clearly thermophilic, there tend to be relatively small differences between Teq and growth temperature (7.3⫾6.6°C); however, for psychrophiles, psychrotrophs, and low-temperature mesophiles (growth temperature ⱕ31°C), they are larger (27.8⫾15.1°). A nonparametric t test between the Teq and growth temperature differences of thermophilic and nonthermophilic enzymes reveals that the two groups are significantly different (P value⫽0.008) in this respect. A previous study using kcat as a measure of temperature optima also suggested that enzymes tend to have optimal activity at temperatures higher than their hosts’ living temperature, and the higher the growth temperature of the organism the narrower the gap becomes (21). There are several possible reasons for this. There may be a poor match between the environmental temperature of thermophiles and their laboratory growth temperature optima; thermophiles are often isolated from sources that are appreciably hotter than their determined growth temperatures. Alternatively, it might be argued that the origin of the current mesophilic microbial life is thermophilic, i.e., that microbial evolution has proceeded down-temperature from a (hyper)thermophilic last common ancestor (22–24), and that the evolution of Teq lags behind optimal growth temperature. Furthermore, depending on how crucial their roles may be in the cell, different enzymes might evolve at different speeds to adapt their Teq to the environment, thereby resulting in the variation in Teq among enzymes from one organism (e.g., for Bos taurus, Table 1). Conclusions based on weak correlations must be speculative. Although its P value is slightly over the adopted threshold of 0.05, the correlation between ⌬G‡cat and ⌬G‡inact may suggest that ⌬G‡cat is higher for more stable proteins and thus support various studies (21, 25, 26) that suggested that a change in activation energy (⌬G‡cat) might be important to the temperature-adaptation of psychrophilic enzymes. To investiTEMPERATURE DEPENDENCE OF ENZYME ACTIVITY
gate this idea, correlation analysis was performed with psychrophilic and psychrotrophic enzymes (growth temperatureⱕ20°C, not including plant enzymes) excluded; the correlation between ⌬G‡inact and ⌬G‡cat was found to deteriorate significantly (P value⬎0.1), tending to confirm that a lower ⌬G‡cat might indeed be one of the main adaptations for psychrophilic enzymes, while nonpsychrophilic enzymes take on other means of adaptation. The scatter plots in Table 2 are consistent with the correlations described. Growth temperature In general, any correlations with growth temperature must be regarded as tentative since determining environmental growth temperature from laboratory-determined optimal growth temperatures of bacteria is problematic, partly because the temperature of most microbial environments varies with time, including that of Escherichia coli, for example (27, 28), and partly because the laboratory environment on which these are usually based may not represent environmental conditions. In the absence of better measures, however, we have generally adopted here the published data for organisms involved in this study (see Supplemental Data). Individual enzymes The DHFR from the strict psychrophile Moritella profunda, isolated from deep Atlantic sediments (29, 30), is a surprisingly stable enzyme with a concomitantly high Teq, while its ⌬Heq (104 kJ䡠mol⫺1) is significantly lower than the 25th percentile (115 kJ䡠mol⫺1), suggesting eurythermal behavior. We speculate that its low temperature activity has been achieved by evolving a lower ⌬Heq into an enzyme with an incidentally high Teq, thus conferring eurythermalism. Part of the lower half (below Topt) of the enzyme activity curve is influenced by ⌬Heq, but the effect of ⌬G‡cat may dominate, as stated above. The alkaline phosphatase from isolate ASB HK47 (Fig. 2A) gives another illustration of the relative effects of the four parameters on enzyme activity over a range of temperatures. For many enzymes (e.g., B. cereus DHFR, Fig. 2B), the denaturation rate is such that it will play the major role in controlling enzyme activity over timescales important to the organism (over minutes rather than seconds) at temperatures above Teq. However, for the ASB HK47 alkaline phosphatase, it is evident that the lower rate of denaturation above Teq leads to a situation where activity will be dominated by Teq and ⌬Heq over significant timescales. To achieve temperature adaptation, evolution imposes a strong selection force for thermophilic enzymes to have both high thermal stability and a high Teq. High thermal stability has been thought to be achieved at the expense of flexibility to explain why thermophilic enzymes generally exhibit relatively little activity at lower temperatures (31). However, the rigidity required for 1939
stability is global, while the flexibility required for catalysis is likely to be local (32), as demonstrated by some exceptions (33, 34). It has been suggested that a combination of local flexibility at the active site and high overall stability (35, 36), is behind such phenomena. In other words, enzymes active at low temperatures do not necessarily have to be thermally unstable (37, 38). Recent reports also suggested that the higher flexibility of psychrophilic enzymes is local and on the microsecond to millisecond timescale (39), while their global flexibility may not be necessarily higher (38). In addition, random and directed mutagenesis studies (37, 40, 41) have shown that it is possible to engineer an enzyme so that it retains low-temperature activity while gaining thermostability.
grant from the New Zealand Marsden Fund. We thank Y. Xu at Free University of Brussels for M. profunda DHFR, C. Monk for providing purified Caldicellulosiruptor saccharolyticus -GLU, D. Clement for providing purified Geobacillus stearothermophilus DHFR, and J. Klinman at UC Berkeley for the gift of the G. stearothermophilus DHFR clone. We also thank Dr. Ray Littler at the University of Waikato for some assistance with the statistical analysis. We are grateful to O. Planckaert and A.-C. Tsuei for help with data collection.
REFERENCES 1. 2.
CONCLUSION
3.
The correlations observed here do not necessarily denote a causative relationship. The finding that Teq correlates best with growth temperature does not necessarily mean that it is the parameter most directly selectable by evolution, although it does provide insight into the effect of temperature on the evolution of enzymes. It should also be noted that the strong correlation between Teq and growth temperature means that the correlation is likely to be reproducible with any selection of enzymes of similar number and diversity to this study. ⌬Heq is a new quantitative measure to characterize eurythermalism at an enzymatic level. The new parameters (i.e., Teq and ⌬Heq), combined with established parameters (i.e., ⌬G‡cat and ⌬G‡inact), define fully how temperature influences enzyme activity, while the former provide new tools to describe and investigate this influence. Much of the data presented here comes from bacterial sources, although results from correlation analysis remain similar when data from eukaryotic sources are excluded (data not shown). Although the 21 sets of data from discrete enzymes presented here are a relatively small data set, the main correlations are robust and generally consistent with findings or proposals based on other data; they also provide new targets and perspectives for molecular and structural studies. The development and verification of the Equilibrium Model have provided additional and quantitative measures of the thermal behavior of enzymes (1, 15) that are essential for describing the effect of temperature on enzyme activity and useful parameters for measuring the temperature adaptation of enzymes. Additionally, the results here show the Equilibrium Model’s potential usefulness as a tool to investigate the relationship between enzyme thermal properties and the influence of temperature on the physiology and evolution of the host organism. This work was partly supported by a grant from the National Science Foundation (Biocomplexity 0120648) and a 1940
Vol. 21
June 2007
4. 5.
6.
7. 8.
9. 10. 11.
12. 13. 14.
15.
16. 17.
Daniel, R. M., Danson, M. J., and Eisenthal, R. (2001) The temperature optima of enzymes: a new perspective on an old phenomenon. Trends Biochem. Sci. 26, 223–225 Thomas, T. M., and Scopes, R. K. (1998) The effects of temperature on the kinetics and stability of mesophilic and thermophilic 3-phosphoglycerate kinases. Biochem. J. 330, 1087–1095 Gerike, U., Danson, M. J., Russell, N. J., and Hough, D. W. (1997) Sequencing and expression of the gene encoding a cold-active citrate synthase from an Antarctic bacterium, strain DS2–3R. Eur. J. Biochem. 248, 49 –57 Notomista, E., Catanzano, F., Graziano, G., Dal Piaz, F., Barone, G., D’Alessio, G., and Di Donato, A. (2000) Onconase: an unusually stable protein. Biochemistry 39, 8711– 8718 Padmanabhan, S., Zhou, K., Chu, C. Y., Lim, R. W., and Lim, L. W. (2001) Overexpression, biophysical characterization, and crystallization of ribonuclease I from Escherichia coli, a broadspecificity enzyme in the RNase T2 family. Arch. Biochem. Biophys. 390, 42–50 Noda, L., and Kuby, S. A. (1957) Adenosine triphosphateadenosine monophosphate transphosphorylase (myokinase). I. Isolation of the crystalline enzyme from rabbit skeletal muscle. J. Biol. Chem. 226, 541–549 Tsuboi, K. K., and Chervenka, C. H. (1975) Adenylate kinase of human erythrocyte. Isolation and properties of the predominant inherited form. J. Biol. Chem. 250, 132–140 Diaz, A., Rangel, P., Montes de Oca, Y., Lledias, F., and Hansberg, W. (2001) Molecular and kinetic study of catalase-1, a durable large catalase of Neurospora crassa. Free. Radic. Biol. Med. 31, 1323–1333 Calera, J. A., Sanchez-Weatherby, J., Lopez-Medrano, R., and Leal, F. (2000) Distinctive properties of the catalase B of Aspergillus nidulans. FEBS Lett. 475, 117–120 Wasserman, B. P., and Hultin, H. O. (1981) Effect of deglycosylation on the stability of Aspergillus niger catalase. Arch. Biochem. Biophys. 212, 385–392 Lopez-Medrano, R., Ovejero, M. C., Calera, J. A., Puente, P., and Leal, F. (1995) An immunodominant 90-kilodalton Aspergillus fumigatus antigen is the subunit of a catalase. Infect. Immun. 63, 4774 – 4780 Switala, J., O’Neil, J. O., and Loewen, P. C. (1999) Catalase HPII from Escherichia coli exhibits enhanced resistance to denaturation. Biochemistry 38, 3895–3901 Daniel, R. M., Cowan, D. A., Morgan, H. W., and Curran, M. P. (1982) A correlation between protein thermostability and resistance to proteolysis. Biochem. J. 207, 641– 644 Owusu, R. K., and Cowan, D. A. (1989) Correlation between microbial protein thermostability and resistance to denaturation in aqueous: organic solvent two-phase systems. Enzyme Microbial Techn. 11, 568 Peterson, M. E., Eisenthal, R., Danson, M. J., Spence, A., and Daniel, R. M. (2004) A new intrinsic thermal parameter for enzymes reveals true temperature optima. J. Biol. Chem. 279, 20717–20722 Eisenthal, R., Peterson, M. E., Daniel, R. M., Danson, M. J. (2006) The thermal behaviour of enzymes: implications for biotechnology. Trends Bio/Tech. 24, 289 –292 Peterson, M. E., Daniel, R. M., Danson, M. J., and Eisenthal, R. (2007) The dependence of enzyme activity on temperature: determination and validation of parameters. Biochem. J. 402, 331–337.
The FASEB Journal
LEE ET AL.
18. 19. 20. 21.
22. 23. 24. 25.
26.
27. 28. 29.
30.
31.
Peterson, M. E. (2005) Evidence for a third thermal parameter of enzymes. PhD Thesis: University of Waikato, Hamilton, New Zealand Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248 –254 StatSoft Inc (2005) STATISTICA, version 7.1. p. www.statsoft. com, Tulsa, OK 74104 U. S. A. Georlette, D., Blaise, V., Collins, T., D’Amico, S., Gratia, E., Hoyoux, A., Marx, J. C., Sonan, G., Feller, G., and Gerday, C. (2004) Some like it cold: biocatalysis at low temperatures. FEMS Microbiol. Rev. 28, 25– 42 Schwartzman, D. W., and Lineweaver, C. H. (2004) The hyperthermophilic origin of life revisited. Biochem. Soc. Trans. 32, 168 –171 Pace, N. R. (1997) A molecular view of microbial diversity and the biosphere. Science 276, 734 –740 Wachtershauser, G. (1998) Thermophiles: the Keys to Molecular Evolution and the Origin of Life? (Wiegel, K., Adams, M., ed) pp. 47–57, Taylor and Francis, Philadelphia D’Amico, S., Claverie, P., Collins, T., Georlette, D., Gratia, E., Hoyoux, A., Meuwis, M. A., Feller, G., and Gerday, C. (2002) Molecular basis of cold adaptation. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 357, 917–925 Johns, G. C., and Somero, G. N. (2004) Evolutionary convergence in adaptation of proteins to temperature: A4-lactate dehydrogenases of Pacific damselfishes (Chromis spp.). Mol. Biol. Evol. 21, 314 –320 Flint, K. P. (1987) The long-term survival of Escherichia coli in river water. J. Appl. Bacteriol. 63, 261–270 Hartl, D. L., and Dykhuizen, D. E. (1984) The. population. genetics. of. Escherichia. coli. Annu. Rev. Genet. 18, 31– 68 Xu, Y., Nogi, Y., Kato, C., Liang, Z., Ruger, H. J., De Kegel, D., and Glansdorff, N. (2003) Moritella profunda sp. nov., and Moritella abyssi sp. nov., two psychropiezophilic organisms isolated from deep Atlantic sediments. Int. J. Syst. Evol. Microbiol. 53, 533–538 Xu, Y., Feller, G., Gerday, C., and Glansdorff, N. Moritella cold-active dihydrofolate reductase: are there natural limits to optimization of catalytic efficiency at low temperature? (2003) J. Bacteriol. 185, 5519 –5526 Ichikawa, J. K., and Clarke, S. (1998) A highly active protein repair enzyme from an extreme thermophile: the L-isoaspartyl methyltransferase from Thermotoga maritima. Arch. Biochem. Biophys. 358, 222–231
TEMPERATURE DEPENDENCE OF ENZYME ACTIVITY
32. 33.
34.
35.
36. 37. 38.
39.
40.
41.
Daniel, R. M., Dunn, R. V., Finney, J. L., and Smith, J. C. (2003) The role of dynamics in enzyme activity. Annu. Rev. Biophys. Biomol. Struct. 32, 69 –92 Merz, A., Knochel, T., Jansonius, J. N., and Kirschner, K. (1999) The hyperthermostable indoleglycerol phosphate synthase from Thermotoga maritima is destabilized by mutational disruption of two solvent-exposed salt bridges. J. Mol. Biol. 288, 753–763 Pappenberger, G., Schurig, H., and Jaenicke, R. (1997) Disruption of an ionic network leads to accelerated thermal denaturation of D-glyceraldehyde-3-phosphate dehydrogenase from the hyperthermophilic bacterium Thermotoga maritima. J. Mol. Biol. 274, 676 Aguilar, C. F., Sanderson, I., Moracci, M., Ciaramella, M., Nucci, R., Rossi, M., and Pearl, L. H. (1997) Crystal structure of the beta-glycosidase from the hyperthermophilic archeon Sulfolobus solfataricus: resilience as a key factor in thermostability. J. Mol. Biol. 271, 789 – 802 Daniel, R. M., Dines, M., and Petach, H. H. (1996) The denaturation and degradation of stable enzymes at high temperatures. Biochem. J. 317, 1–11 Van den Burg, B., Vriend, G., Veltman, O. R., Venema, G., and Eijsink, V. G. (1998) Engineering an enzyme to resist boiling. Proc. Natl. Acad. Sci. U. S. A. 95, 2056 –2060 Svingor, A., Kardos, J., Hajdu, I., Nemeth, A., and Zavodszky, P. (2001) A better enzyme to cope with cold. Comparative flexibility studies on psychrotrophic, mesophilic, and thermophilic IPMDHs. J. Biol. Chem. 276, 28121–28125 Georlette, D., Damien, B., Blaise, V., Depiereux, E., Uversky, V. N., Gerday, C., and Feller, G. (2003) Structural and functional adaptations to extreme temperatures in psychrophilic, mesophilic, and thermophilic DNA ligases. J. Biol. Chem. 278, 37015–37023 Gershenson, A., Schauerte, J. A., Giver, L., and Arnold, F. H. (2000) Tryptophan phosphorescence study of enzyme flexibility and unfolding in laboratory-evolved thermostable esterases. Biochemistry 39, 4658 – 4665 Miyazaki, K., Wintrode, P. L., Grayling, R. A., Rubingh, D. N., and Arnold, F. H. (2000) Directed evolution study of temperature adaptation in a psychrophilic enzyme. J. Mol. Biol. 297, 1015–1026 Received for publication September 12, 2006. Accepted for publication January 18, 2007.
1941
